hadoop-hdfs-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Nitin Pawar <nitinpawar...@gmail.com>
Subject Re: HDFS Quota Error
Date Thu, 22 May 2014 17:49:21 GMT
based on your table file format along with table definition  and what kind
of query you are running on that data will decide on how many files need to
created. These files are created as temporary output from maps till
reducers consume them.

you can control how many files hive's job should create on run time by
setting
set hive.merge.mapfiles=true;
set hive.exec.max.dynamic.partitions.pernode=10000;
set hive.exec.max.dynamic.partitions=20000; # check what number you want to
set this to based on your machine configs

set hive.exec.max.created.files=200000;

Also if your table has lots of small files then change the input file
format by setting
set hive.input.format=org.apache.hadoop.hive.ql.io.CombineHiveInputFormat;


But this will also depend on what size disk you have and what's your base
filesystem type.
also do not forget to set the ulimit to unlimited.
If have reset ulimit then you will need to restart your hadoop cluster.

wait for some experts from dev forum to give more insights on this


On Thu, May 22, 2014 at 3:53 PM, Natarajan, Prabakaran 1. (NSN -
IN/Bangalore) <prabakaran.1.natarajan@nsn.com> wrote:

>  Hi,
>
>
>
> Thanks.
>
>
>
> Inode is 100% in the disk where it mounted to the directly
> /var/local/hadoop (its not temp, but hadoops working or cache directory).
> This happens when we run aggregation query in hive.  Looks like hive query
> (map-red) create many small files.
>
>
>
> How to control this? What are those files?
>
>
>
> *Thanks and Regards*
>
> Prabakaran.N  aka NP
>
> nsn, Bangalore
>
> *When "I" is replaced by "We" - even Illness becomes "Wellness"*
>
>
>
>
>
> *From:* ext Nitin Pawar [mailto:nitinpawar432@gmail.com]
> *Sent:* Thursday, May 22, 2014 3:07 PM
>
> *To:* user@hadoop.apache.org
> *Subject:* Re: HDFS Quota Error
>
>
>
> That means there are some or a process which are creating tons of small
> files and leaving it there when the work completed.
>
>
>
> To free up inode space you will need to delete the files.
>
> I do not think there is any other way.
>
>
>
> Check in your /tmp folder, how many files are there and if any process is
> leaving tmp files behind.
>
>
>
> On Thu, May 22, 2014 at 2:54 PM, Natarajan, Prabakaran 1. (NSN -
> IN/Bangalore) <prabakaran.1.natarajan@nsn.com> wrote:
>
> Just noted that inode is 100%.  Any better solutions to solve this?
>
>
>
> *Thanks and Regards*
>
> Prabakaran.N  aka NP
>
> nsn, Bangalore
>
> *When "I" is replaced by "We" - even Illness becomes "Wellness"*
>
>
>
>
>
> *From:* ext Natarajan, Prabakaran 1. (NSN - IN/Bangalore) [mailto:
> prabakaran.1.natarajan@nsn.com]
> *Sent:* Thursday, May 22, 2014 2:37 PM
> *To:* user@hadoop.apache.org
> *Subject:* RE: HDFS Quota Error
>
>
>
> Thanks for your reply.  But all the datanote disk has more than 50% space
> empty
>
>
>
> *Thanks and Regards*
>
> Prabakaran.N  aka NP
>
> nsn, Bangalore
>
> *When "I" is replaced by "We" - even Illness becomes "Wellness"*
>
>
>
>
>
> *From:* ext Nitin Pawar [mailto:nitinpawar432@gmail.com<nitinpawar432@gmail.com>]
>
> *Sent:* Thursday, May 22, 2014 12:56 PM
> *To:* user@hadoop.apache.org
> *Subject:* Re: HDFS Quota Error
>
>
>
> no space left on device can also mean that one of your datanode disk is
> full.
>
>
>
> Can you check disk used by each datanode.
>
>
>
> May be you will need to rebalance your replication so that some space is
> made free on this datanode.
>
>
>
> On Thu, May 22, 2014 at 12:34 PM, Natarajan, Prabakaran 1. (NSN -
> IN/Bangalore) <prabakaran.1.natarajan@nsn.com> wrote:
>
> Hi
>
>
>
> When I run a query in Hive, I get below exception.  I noticed the error
> “No space left on device”.
>
>
>
> Then I did “hadoop fs -count -q /var/local/hadoop” – which gave below
> output
>
>
>
> none             inf            none             inf           69
> 275          288034318 hdfs://nnode:54310/var/local/hadoop
>
>
>
> Why I am getting none and inf for space and remaining space quota?  Is
> this meaning is unlimited space or is there is any space left?
>
>
>
>
>
> I tried “hadoop dfsadmin -setSpaceQuota 100G /var/local/hadoop”  --  Not
> sure 100G is correct or not?  How much I need to set and how to calculate
> this?
>
>
>
> After setting 100G , I get the below output  for “hadoop fs -count -q
> /var/local/hadoop”
>
>
>
> none             inf    107374182400    104408308039           73
> 286          297777777 hdfs://nnode:54310/var/local/hadoop
>
>
>
>
>
> I have to wait to see whether 100G is going to give me an exception or
> not….
>
>
>
>
>
> --------------
>
>
>
>
>
> 2014-05-22 10:48:43,585 ERROR
> org.apache.hadoop.security.UserGroupInformation: PriviledgedActionException
> as:hadoop cause:java.io.FileNotFoundException:
> /var/local/hadoop/cache/mapred/local/taskTracker/hadoop/jobcache/job_201405211712_0625/jars/org/iq80/snappy/SnappyCompressor.class
> (No space left on device)
>
> 2014-05-22 10:48:43,585 WARN org.apache.hadoop.mapred.TaskTracker: Error
> initializing attempt_201405211712_0625_r_000001_2:
>
> java.io.FileNotFoundException:
> /var/local/hadoop/cache/mapred/local/taskTracker/hadoop/jobcache/job_201405211712_0625/jars/org/iq80/snappy/SnappyCompressor.class
> *(No space left on device)*
>
>         at java.io.FileOutputStream.open(Native Method)
>
>         at java.io.FileOutputStream.<init>(FileOutputStream.java:221)
>
>         at java.io.FileOutputStream.<init>(FileOutputStream.java:171)
>
>         at org.apache.hadoop.util.RunJar.unJar(RunJar.java:51)
>
>         at
> org.apache.hadoop.mapred.JobLocalizer.localizeJobJarFile(JobLocalizer.java:277)
>
>         at
> org.apache.hadoop.mapred.JobLocalizer.localizeJobFiles(JobLocalizer.java:377)
>
>         at
> org.apache.hadoop.mapred.JobLocalizer.localizeJobFiles(JobLocalizer.java:367)
>
>         at
> org.apache.hadoop.mapred.DefaultTaskController.initializeJob(DefaultTaskController.java:202)
>
>         at
> org.apache.hadoop.mapred.TaskTracker$4.run(TaskTracker.java:1228)
>
>         at java.security.AccessController.doPrivileged(Native Method)
>
>         at javax.security.auth.Subject.doAs(Subject.java:415)
>
>         at
> org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1121)
>
>         at
> org.apache.hadoop.mapred.TaskTracker.initializeJob(TaskTracker.java:1203)
>
>         at
> org.apache.hadoop.mapred.TaskTracker.localizeJob(TaskTracker.java:1118)
>
>         at
> org.apache.hadoop.mapred.TaskTracker$5.run(TaskTracker.java:2430)
>
>         at java.lang.Thread.run(Thread.java:744)
>
>
>
>
>
> *Thanks and Regards*
>
> Prabakaran.N  aka NP
>
> nsn, Bangalore
>
> *When "I" is replaced by "We" - even Illness becomes "Wellness"*
>
>
>
>
>
>
>
>
>
>
>
>
>
> --
> Nitin Pawar
>
>
>
>
>
> --
> Nitin Pawar
>



-- 
Nitin Pawar

Mime
View raw message